Five Tips for Better Demand Planning and Forecasting

Oct. 12, 2007
Start with the premise that everything will not go exactly as planned.

Yogi Berra once said, "It's tough to make predictions, especially about the future." While there's no magic formula for forecasting, there are several steps that companies can take to mitigate uncertainty and improve their demand planning capabilities. Knowing these steps is critical in helping companies avoid issues further down the supply chain, says Randy Strang, vice president of global solutions and implementation-retail with UPS Supply Chain Solutions. Strang offers the following tips to improving forecast accuracy.

1. One size does not fit all.

Forecasting needs and challenges vary widely, depending on a company's business model, size, geographic location and industry sector, among other factors. Even within a company, there are several different situations that require some form of forecasting for management control. Before tackling the demand planning process, manufacturing companies should clearly define their forecasting problems:

  • Are you setting stocking policies or setting financial guidance?
  • Do you have sporadic/lumpy demand or continuous volume?
  • Are you looking at short-term or long-term financial projections?
  • What are your allowable tolerances?

2. Understand the drivers of uncertainty.

There are numerous factors that can impact a company's ability to forecast accurately. Knowing these factors in advance helps companies plan ahead. Manufacturers should ask themselves:

  • How consistent is your demand?
  • What are the factors that influence the variable, or variables being forecast?
  • What level of supply chain visibility do you have?
  • How reliable are your modes of transport?
  • Do you have access to multiple modes of transport?

3. Keep it simple.

Complex is not always better when it comes to demand planning. The forecasting method and tools that are right for a particular company depends on the level of support they need (see #1) and the level of data they have. The right solution could be a statistical approach or a consensus approach, a stand-alone tool or an enterprisewide solution. Common forecasting methods include:

  • Time series methods look at historical data and project forward.
  • Regression methods examine previous/historical averages and outcomes and hypothesize relationships among variables.
  • Heuristic methods leverage the experience and expertise of company leaders.
  • Consensus approach methods involve the right players across an organization.

4. Prepare for change.

Supply Chain Wiki

The Warehousing Education and Research Council (WERC) has launched a wiki aimed at supply chain professionals. The WERCipedia is designed to act as an interactive online glossary, with a searchable collection of terminology, acronyms, abbreviations and other information. Users will be able to add new terms as well as modify definitions of existing ones. You can check it out at

Understanding the core competencies that a company needs in order to succeed at demand planning is critical. Before they embark on implementing or changing the demand planning process, companies should:
  • Ensure that you have the right skill sets for your needs.
  • Put the right training programs in place.
  • Engage in sales and operations planning where cross-functional planning is required.
  • Define methods for measuring performance.

5. Expect the unexpected.

The only guarantee in forecasting is that everything will not go exactly as planned. This is why having defined alternatives or backup plans (or what UPS likes to call a demand- responsive and flexible supply chain) is crucial. Companies should:

  • Ensure that you have the flexibility to quickly obtain alternate supplies from the field and a time-sensitive service capability to deliver these.
  • Look at time-definite transportation services and options for shortening lead times and making them less variable.
  • Ensure that you are working with carriers/vendors/partners that have flexible business models.
  • Plan ahead -- for every scenario.

See Also

Popular Sponsored Recommendations

Voice your opinion!

To join the conversation, and become an exclusive member of IndustryWeek, create an account today!